Kim et al. (2026) Insights into uncertainties in future drought analysis using hydrological simulation model

Identification

• Journal: Hydrology and earth system sciences
• Year: 2026
• Date: 2026-01-15
• Authors: Jin Hyuck Kim, Eun-Sung Chung
• DOI: 10.5194/hess-30-227-2026

Research Groups

• Department of Civil Engineering, Chungnam National University, South Korea
• Faculty of Civil Engineering, Seoul National University of Science and Technology, South Korea

Short Summary

This study quantifies uncertainties in future runoff and hydrological drought projections, revealing that while General Circulation Models (GCMs) are the dominant source, hydrological model calibration contributes 3.9 %–9.8 % to runoff uncertainty (especially in low flow periods) and 2.7 % to drought uncertainty, with performance significantly influenced by the hydrological conditions of the calibration data.

Objective

• Quantify uncertainties in future runoff projection and hydrological drought analysis stemming from future climate data (GCMs, Shared Socioeconomic Pathways - SSPs) and hydrological model calibration parameters (data length and hydrological conditions).

Study Configuration

• Spatial Scale: Four dam basins in South Korea: Andong (AD), Chungju (CJ), Habcheon (HCH), and Seomjingang (SJ). Basin areas range from 763 square kilometers to 6648 square kilometers.
• Temporal Scale:
  • Historical period for model calibration and validation: 1980–2023.
  • Future periods: Near Future (NF) and Distant Future (DF), analyzed across decades (e.g., 2040s-2090s for drought).
  • Hydrological model calibration data lengths: 1 to 20 years.

Methodology and Data

• Models used:
  • Hydrological model: Soil and Water Assessment Tool (SWAT)
  • Statistical analysis: Analysis of Variance (ANOVA)
  • Drought index: Streamflow Drought Index (SDI)
  • Parameter optimization: SUFI-2 algorithm (within R-SWAT)
  • Bias correction: Quantile mapping (linear parametric transformation)
  • Spatial interpolation: Inverse Distance Weighting (IDW)
• Data sources:
  • Future climate data: Outputs from 20 General Circulation Models (GCMs) from CMIP6 under three Shared Socioeconomic Pathway (SSP) scenarios (SSP2-4.5, SSP3-7.0, and SSP5-8.5).
  • Observed meteorological data: Korea Meteorological Administration (KMA) weather stations (for GCM bias correction).
  • Observed hydrological data: Daily dam inflow records (1980–2023) for model calibration and validation.
  • Topographic data for the dam basins.

Main Results

• Hydrological model parameters calibrated using dry period data showed an average of 11.4 % higher performance when validated under dry conditions and 6.1 % higher performance under normal conditions, compared to parameters calibrated with wet period data. Conversely, wet period calibrated parameters showed 5.1 % higher performance in wet conditions.
• The optimal calibration data length for best performance varied by basin, and longer calibration periods did not consistently guarantee improved model performance.
• General Circulation Models (GCMs) were identified as the dominant source of total uncertainty in future runoff projections, consistently contributing over 60 % on average.
• The uncertainty contribution from hydrological model calibration (due to data length and hydrological conditions) in estimating future runoff ranged from 3.9 % to 9.8 %, being particularly significant during low runoff periods (e.g., spring and winter).
• The uncertainty contribution from hydrological model calibration in future hydrological drought analysis averaged 2.7 %, which is lower than that observed for future runoff projections and exhibited different basin-specific patterns.
• Interactions between GCMs and SSPs, and between GCMs and hydrological model calibration factors (hydrological conditions and period length), were found to be statistically significant, indicating a complex interplay of uncertainty sources.

Contributions

• Quantifies the cascade of uncertainties in future runoff and hydrological drought analysis, specifically focusing on the impact of hydrological model calibration (data length and hydrological conditions) in addition to GCMs and SSPs.
• Demonstrates that hydrological conditions during calibration have a greater impact on uncertainty than calibration data length.
• Highlights the necessity for separate analyses of uncertainties in future runoff projection and hydrological drought analysis due to differing characteristics and contributions.
• Provides insights into basin-specific sensitivities to calibration choices, particularly for basins with lower precipitation.

Funding

• National Research Foundation of Korea (grant no. 2021R1A2C200569915)

Citation

@article{Kim2026Insights,
  author = {Kim, Jin Hyuck and Chung, Eun-Sung},
  title = {Insights into uncertainties in future drought analysis using hydrological simulation model},
  journal = {Hydrology and earth system sciences},
  year = {2026},
  doi = {10.5194/hess-30-227-2026},
  url = {https://doi.org/10.5194/hess-30-227-2026}
}